Hello ​@hogkai ,

Regarding your first question: in order for the AI to respond to the user's initial message, you should send the last input text to the AI block. You can include something like this in your prompt:  
"This is the message the user sent: '{{last_input_text}}'. Continue the conversation."

As for your second question about combining both flows, I’m not sure that would be the best approach. That’s because your ad flow requires a block with at least one button. If you merge everything into the ad trigger, people who message you directly (not through the ad) won’t receive the same content.

Regarding the rest of the automation after the AI step, it's worth checking if it's not progressing because the AI is waiting for a user reply that never comes. From what I can see, your flow seems fine — in fact, the second part has already been sent to several users. So, it’s possible they are stuck in the AI step. If that’s the case, here’s something you can try:

- Tag users as soon as they enter the AI step — for example, with a tag like "Chatting with AI".
- Then set up an action to remove the tag once they leave the AI step. This way, you can easily identify who is stuck.

At the same time, you can create another flow based on the user’s last interaction.  
For example: if it's been X hours since their last message and they still have the "Chatting with AI" tag, it likely means they’re stuck in the AI step. In that case, send a message to re-engage them.
Let me know if I was able to help you.

“As for your second question about combining both flows, I’m not sure that would be the best approach. That’s because your ad flow requires a block with at least one button. If you merge everything into the ad trigger, people who message you directly (not through the ad) won’t receive the same content.”→ For this point, I think it is possible to have two triggers (user send message & click ads) at one flow

You meant sth like this. I quite do not understand your last idea with another flow.

Below is a screenshot to better explain my idea: tagging all contacts currently in the AI stage and sending a reactivation message to those who remain inactive.

1. Create a custom *date and time* field to record when the contact entered this flow. Then, open that field’s settings as shown in the screenshot, and configure an *offset*, meaning how long after the entry point you want to send a message. For example, if the contact entered the AI stage at 5:00 AM and you set an offset of +6 hours, the field will save the value 11:00 AM.

2. In the second step, add an action block that assigns a TAG to the contact. This ensures that 100% of the contacts who start a conversation with the AI will receive this TAG.

3. Right after they leave the AI stage, the TAG will be removed.

You can see in the screenshot below how this will look. This step is essentially logging a timestamp to check whether the lead is still in the AI stage and tagging each lead so you can target only the right ones.

The next step is to create a separate flow that is triggered based on date and time. This date and time comes from the custom field that saved when the contact started the AI conversation, plus the number of hours you define. Be sure to add a condition: the message should only be sent if the TAG is still attached to the contact. See the screenshot of this part below:

 

I have applied all your suggestions in my flow. There are some wonderings:
“Configure an *offset*, meaning how long after the entry point you want to send a message.”

1. This means after offset time, the AI Step would send message to a contact or what?

2. I do not understand the setting in highlight of this picture below:

Hello ​@hogkai ,

This suggested flow is intended to help reactivate customers who got stuck at the AI step for some reason — for example, if they did not complete the interaction with the AI and therefore did not progress to the next stage. After the configured time has passed, the message sent will not be from the AI, but rather a fixed, pre-defined message, simply meant to re-engage the customer.

Regarding your question about the field configuration: this flow will be triggered at the time saved in that field. For example:

If the user started the flow at 4:00 a.m. and you set an offset of +10 hours, then the personalized field will contain the time 2:00 p.m. This means that the flow will be sent to the customer exactly at 2:00 p.m.

Does the offset time setting affect AI Step response time?

This does not affect the AI's response time in any way. Just a reminder that the AI already sends an automatic follow-up message, if I’m not mistaken, when the user hasn’t replied for about 30 minutes.

So, this flow only makes sense for longer periods — for instance, when the AI has already sent the initial message and the follow-up, and 5 or 10 hours have passed without any response from the person.

I applied your suggested flows.

Now the pre-defined message I want to send to contact after AI Step has been sent as soon as AI Step happen 🥲
Can you help me to fix it?

Sure, this needs to be adjusted.

Please keep the following in mind: when you send a message created in this new flow, the AI step will still be active. This means the AI will not be aware of the message you sent in this new flow, since it is not part of the AI process. As a result, the user might ask a question that doesn't make sense, because the AI won’t have the context of your message.

Given that, you should consider whether it's better to go ahead and send your main final message — with the image — or just send a simple reactivation message like: “Are you still there? Can we continue our conversation?” This kind of message won’t affect the AI's interaction.

That said, if you want to disable the message that’s set to be sent after the AI step when you send the follow-up message, you’ll need to use TAGs for that.

At the beginning of the new flow, add a TAG like “Reactivation message sent”.
Then, in the original flow (after the AI step), add a condition: if the user has the “Reactivation message sent” TAG, do not send the final message — stop the flow. This way, if the new flow has already been sent, the final message won't be sent again.

Your point: “Then set up an action to remove the tag once they leave the AI step. This way, you can easily identify who is stuck.”

I still do not know what conditions make AI Step over and move to remove TAG action?

The only condition that makes the AI move forward — and consequently proceed to the next step and remove the TAG — is if the user interacts with it. If the user doesn't respond, it will remain idle.

I just want to my CTA Message be sent after last interaction of users with AI Chat. Can I use condition to make it, can you suggest me ways to do it. Thank you!

When the user completes the interaction with the AI, they automatically proceed to the next step in the flow, where the CTA message is sent. No specific condition is required for this to happen.

If some users are not moving forward, observe how the interaction between the AI and the user is taking place. You can influence this behavior—whether the transition happens more quickly or slowly—through your prompt design and the tasks you assign to the AI.

When the AI understands that it has completed the task you set, it ends the interaction, and the user automatically advances to receive the CTA.

You can modify and test your prompt settings by engaging with the AI and checking whether its behavior aligns with what you expect.